1

Retail and Wholesale Trade

E-Commerce:
What the Numbers Really Mean

*



Eric Brousseau**

A review of quantitative research on electronic commerce (e-commerce) and its impact
reveals major gaps in our knowledge about current changes. The recent introduction of
some online applications and the rapid pace of change are not the only explanations. The
great store placed in the myth that perfect markets will be achieved through digital
networks and the emphasis on online ordering explain many of the obstacles to proper
understanding and measurement of the e-commerce phenomenon.
Three important facts need to be stressed. First of all, the Internet did not beget
e-commerce. E-commerce developed from other technologies and several forms of
e-commerce will continue to rely on other networks than the Internet. Secondly, many
electronic systems are not designed to handle all of the operations involved in a
commercial transaction. Digital networks are just one of the channels through which
transactions can be conducted in conjunction with other channels. Depending on the
operations that networks are built to handle, different types of e-commerce are being
developed with contrasting effects. And, thirdly, in many cases, digital networks are
used more for differentiating services and targeting specific groups of consumers, rather
than for creating more transparent markets.
Agencies in charge of monitoring e-commerce need to develop more sophisticated
evaluations of how electronic applications are being used and the rationale of the
systems being set up. Such evaluations would provide information on how e-commerce

is being introduced in the various components of the economy.

* Originally published as Commerce lectronique : ce que disent les chiffres et ce quil faudrait savoir, conomie et Statistique,
No. 339-340, 2000 - 9/10.
** Eric Brousseau is a professor at the University of Paris X, Forum & Atom

The names and dates in parentheses refer to the bibliography at the end of the article.
2

ecision-makers enthusiastic talk about the
sheer size and potential of the

new econ-
omy

in general, and e-commerce in particular, is
in striking contrast to the dearth of reliable data
about concrete results. Except for the first
release of official statistics on e-commerce
retail sales in the United States in 2000 (U.S.
Census Bureau of the Department of Com-
merce, 2000), private-sector consulting and
marketing firms produced the vast majority of
the e-commerce data that are now being circu-
lated and even finding their way into official
reports. Such firms as International Data Corpo-
ration (IDC), the Gartner Group, KPMG, For-
rester Research, Jupiter Communication and
Accenture have produced these data using
methodologies that are not always very trans-
parent. The nub of the problem lies in the fact
that most of the available data are forecasts

1

, not
evaluations, and the data from different sources
are not compatible

2

.
The OECD (2000b) cites one of the most blatant
examples: estimates for 1995-1997 vary from
USD 70 billion to USD 24 billion, depending on
the firm providing them (see Table 1). The same
firms estimates made between 1995 and 1998
call for e-commerce to account for an average of
5% of retail sales in 2001 and 2002. These fig-
ures are totally incompatible with the Depart-
ment of Commerces own figures, which put the
share of e-commerce in 2000 retail sales at
about 0.7%, even though the American retail
trade is by far the most digital in the world.
Regardless of the method we use to approach
e-commerce data, we see that all of the discus-
sion, investment strategies and government pol-
icies related to e-commerce are based on very
questionable numbers. Even the bodies that pro-
duce the numbers defend them only half-heart-
edly. They admit that their definitions of e-com-
merce and categories of e-commerce, as well as
their estimation and sampling techniques, differ
and that they are prone to problems and biased.
The key argument in defence of producing such
numbers is that the real issue is to capture the
trend, which shows strong growth in this case.
We could counter that the different bodies esti-
mated growth rates are wildly divergent. For
example, evaluations of the growth of e-com-
merce in Britain show estimated annual growth
rates that vary by a factor of 2 between Optime-
dias estimate of 67% and NOPs estimate of
116%. This means that estimates will diverge by
a factor of 14, ten years down the road (CRIC,


1



2
D

1

Many estimates have been revised downwards or turned out to
be false. For example, the U.S. Department of Commerce noted
that estimates published in 1999 of 1998 online retail sales vol-
ume put the gure at between USD 7 billion and USD 15 billion.
This is to be compared to the gure of USD 4 billion to USD 14
billion estimated in 2000 for 1999 online retail sales volume, and
to the gure of USD 5.3 billion measured by the Department of
Commerce itself. Going by these gures, we see that the
expected explosion of sales volume did not occur and estimated
online sales volume was trimmed in consequence.

2

The French E-Commerce Task Force emphasised that evalua-
tions of B2B (business-to-business) online sales volume in 1998
ranged from USD 3 billion to USD 15 billion (Lorentz, 1998). The
report on the impact of IT by the French General Planning Com-
mission cited discrepancies to the order of one to ten between
rival estimates (Dang NGuyen, 1999). The accuracy of estimates
has hardly improved at all over the last two years. In its 2000
report on the digital economy, the U.S. Department of Com-
merce reported that private-sector estimates of B2C (busi-
ness-to-consumer) sales volume varied widely (see note 1). The
same holds true for estimates of B2B sales volume, which ranged
from USD 634 billion to USD 2,300 billion for 2003 (U.S. Depart-
ment of Commerce, 2000).
Source: OECD, 2000b.

Table 1

The Main E-Commerce Estimates in the United States

In USD millions

Source 1995-1997 2000-2002
IDC 1,000 117,000
INPUT 70 165,000
VeriFone 350 65,000
ActivMedia 24,400 1,522,000
Data Analysis 2,800 217,900
Yankee 850 144,000
E-land 450 10,000
EITO 475 262,000
AEA/AU 200 45,000
Hambrecht & Quest 1,170 23,200
Forrester 8,000 327,000
Morgan Stanley 600 375,000

Mean 725 154,500
3

1999). Therefore, the different firms do not
even agree that we are at the same point on the
S-shaped diffusion curve (since we are dealing
with network technology, see Curien, 2000) and
it is only reasonable to have very serious doubts
about their ability to account even for trends.

A Phenomenon that Takes Many
Forms and Changes Rapidly

he inaccuracy of estimates is easy to
explain. It stems mainly from three over-
lapping phenomena. First of all, we are dealing
with phenomena where quantities and qualities
are both changing very rapidly (see Table 2).
According to the International Telecommunica-
tions Union, the number of Internet users in the
European Union/United States/Japan area
increased from 70 million to 174 million
between 1997 and 1999 (Eurostat, 2001)

3

.
According to the U.S. Department of Com-
merce (1999), Internet traffic volume doubles
every 100 days. The number of secure servers,
which are especially useful for e-commerce,
increased by 95% in the OECD countries
between April 1999 and April 2000 (OECD,
2000a).
This exponential growth comes with a rapid
diversification of use. The growing numbers of
Internet sites are further testimony to diversifi-
cation. The number of sites swelled from 26,000
in 19993 to 5 million in 1999 (U.S. Department
of Commerce, 1999). Even though there is some
overlap, the range of services on offer is much
broader today than it was a few years ago. The
change is obvious in the case of online retailing.
Lucking-Reiley and Spulber (2000) reviewed
all of the marketplace. More than 75% of the 50
or so marketplaces are less than three years old.
In the B2C (business-to-consumer) field, the
public at large is aware of or has heard from the
media about the explosive growth of online
services. These include more or less sophisti-
cated distance-selling techniques in all business
areas or more innovative sites offering new
functions, ranging from auctions and group pur-
chasing to networked games.

3

Given these circumstances, it is difficult to
come up with stable classifications and put
together statistical series, or even to agree on
standardised measurement methods. E-com-
merce is not the only area where this problem is
encountered. The same problem comes up for
all forms of information and communication
technology (ICT) and all its uses, as was empha-
sised in a French General Planning Commission
report (Brousseau and Rallet, 1999). Even
inside firms, it is hard to evaluate the real costs
and the real impact of investment in ICT. Staff
training, maintenance and organisational costs
are often omitted. The impactin terms of jobs,
productivity and competitiveness is very com-
plex, because the technologies involved change
so quickly and affect all business areas. ICT use
is governed by more than one decision-making
circuit and its effectiveness is very dependent
on supporting measures. This means that, even
at the firm level, it is difficult to come up with
reliable figures.
Secondly, we are dealing with very sensitive
information and strategic concerns distort the
T

3

The estimated number of Internet users around the world in
1993 was approximately 3 million (U.S. Department of Com-
merce, 1999).
Source: Eurostat 2000.

Table 2

Changes in E-Commerce Infrastructure 1995-1999

In millions of Units

France EU USA Japan
1995 1999 1995 1999 1995 1999 1995 1999
Number of PCs
Per 100 people
7.8

13

13.0


22

56.2


15

93.2


25

86.3


33

141.0

52

15.1

12

36.3

29

Internet Hosts
Per 100 peoples
0.15

0.3

0.6

1.0

1.8

0.5

8.5


2.3

6.0
2.3
53.1

19.6

0.26
0.2
2.6

2.1

Internet users
Per 100 people
1.0*

1.7*

5.6

9.6

19.3*


5.2

*
55.9


14.9

40.0*


15*

110.0

40.5

11.5*

9.3*

18.3

14.5

Mobile telephone

Per 100 people

1.0

1.8

21.4

36.3

21.1


5.7

146.5


39.1

33.7


12.9

86.0

31.7

1.7

9.3

56.8


45.0

* 1997 gures
4

behaviour of firms with regard to disclosure and
diffusion. We are faced with emerging markets
and uses in a context of technological competi-
tion. Online service providers and the technol-
ogy providers serving them are aware of the
characteristics of an economy with high returns
to adoption and network externalities (David,
1985; Katz and Shapiro, 1985; Arthur, 1989;
Cowan, 1990; Varian and Shapiro, 1999). Pro-
viders have to convince potential adopters or
subscribers that their products and services will
become industry standards or that they will
play a central role in the market. This means that
they are inevitably tempted to exaggerate infor-
mation about their actual or expected market
shares. Furthermore, potential adopters are espe-
cially sensitive to forecasts and beliefs

4

. As
was the case in other ICT applications (Brous-
seau and Rallet, 1999), the uncertainty surround-
ing the impact of new uses, combined with large
real or assumed potential for economies of learn-
ing, created a constant fear of being left behind
by the competition. This fear leads to copying of
competitors adoption choices.

The Myth of Online Ordering

The third phenomenon that explains divergence
between e-commerce evaluations stems from
great uncertainty surrounding the definition and
possibilities of e-commerce. As Rallet (2000)
aptly points out, most of the evaluations are
based on the myth of online ordering. We
could even add that they focus solely on online
ordering over the Internet. This stems from a
biased perception of what ICT contributes to
commercial transactions and a simplistic view
of Internet characteristics. In fact, many opinion
leaders, technology mavens and ordinary citi-
zens have a more or less explicit vision of the
Internet as a global virtual marketplace operat-
ing as a quasi-Walrasian market with not inter-
mediaries between supply and demand

5

. On top
of this, the virtual nature of the electronic mar-
ket is assumed to provide extraordinary savings
on transaction costs.
On the strength of this myth, a lot of research on
e-commerce disregards all forms of commerce
that dont use the Internet, even though they
may use other electronic media. Yet, in func-
tional terms, using e-mail to place an order is
not very different from using a fax or a tele-
phone. Similarly, viewing a catalogue on a web-
site is not very different from consulting a cata-
logue via a videotext system, such as the
Minitel, or even using a telephone to reserve air-
line tickets, for example. Furthermore, conduct-
ing an electronic data interchange (EDI) over
the Internet using WebEDI or over another net-
work, does not change the fact that EDI is
mainly concerned with

ex post

transaction man-
agement, rather than

ex ante

management (see
below). In more general terms, the exclusion of
commerce that does not use the Internet disre-
gards the diversity of uses. Most e-commerce
sites are not anonymous auction markets. Most
sites are more concerned with implementing
new forms of discrimination and intermediation
than they are with creating transparent markets
without intermediaries.

4



5

It is also on the strength of this myth of a virtual
Walrasian market that most statistical work con-
centrates on online ordering. If all of the opera-
tions involved in a transaction are not conducted
online, the innovative effect of the Internet is
assumed to be less significant. If orders are not
placed online, the World Wide Web is no longer
the medium for a fully integrated global market,
it is merely another medium that conventional
operators use for advertising, prospecting for
customers, supplying some additional services
more cheaply, more dynamic catalogue and
price list management, etc.
While it is easy to understand why so many
observers are obsessed with the online ordering
criterion, this position is untenable if we want to
understand the impact of using electronic media
to manage commercial transactions. Yet, this is
a key issue for macroeconomic and microeco-
nomic research. We need to evaluate the impact
of ICT on productivity and market coordination
to understand its macroeconomic impact and
microeconomic research needs to capture firms
strategies and analyse the impact of e-com-
merce on the division of the surplus. It is really
not very relevant to make a distinction between
commerce on the Internet and commerce that
uses other media and to restrict the notion of
electronic commerce to transactions where
orders are placed online when we are trying

to
understand how commerce is using digital tech-
nologies rather than electronic commerce per
se

, as suggested by A. Rallet (2000).

4

In the case of e-commerce, exaggerated demand gures were
further inated by gures presented to investors as the dot-com
stock market bubble grew during 1998 and 1999.

5

A

Walrasian market

is one that centralises the supply and
demand of all of the agents involved in the economy in order to
discover the market-clearing equilibrium prices. If the Internet
were to make the cost of finding information insignificant and
computers did the same thing for the cost of processing informa-
tion, then each agent could have full information about all of the
opportunities and terms of trade in the economy. This would be
the same as knowing the equilibrium price vector.
5

Beyond the Internet:
Other Digital Networks

ather than letting our approach be dictated
by technical criteria, we should look at the
problems in terms of uses. The technical charac-
teristics of the network used are not important;
what counts is how the network is used. This
must be the starting point for any attempt to
measure e-commerce. The approach needs to
focus on services or applications, rather than on
the medium.
The Internet is by no means the sole growth vec-
tor for e-commerce. In France, the Minitel vide-
otext system handles substantial traffic related
to commercial transactions (OECD, 1998). Dig-
ital television cable and satellite networks, as
well as cell telephone networks, are also major
vectors for growth in B2C commerce (see
Box 1). Some 8% of households in the Euro-
pean Union have Internet access, as opposed to
11% in the United States, and 34% of EU house-
holds have cable or satellite television, as
opposed to 36% of American households. In
both the USA and the EU one-third of the pop-
ulation have cell phones and two-thirds have
access to teletex (CRIC, 1999).
Dedicated data transmission networks play an
important role in B2B commerce. Such net-
works are long-established links between most
large and medium-sized firms and they are the
main medium for EDI. According to the Census
Department of the French Ministry of Industry
(Sessi), only 8% of French manufacturing firms
received online orders over the Internet in 1999
(even though some 24% of such firms used the
Internet to diffuse information), yet 44% of
these firms used EDI: 36% for transactions with
partners, 15% for exchanges with tax and
employment authorities and 14% for exchanges
between their various establishments. In France,
business volume conducted via EDI was esti-
mated at FRF 800 billion, as compared to the
estimated volume of FRF 7.3 billion for B2B
business over the Internet (Observatoire du
Commerce et des changes lectroniques,
1999).
Even though there is a gradual shift towards the
Internet, many of the online information mar-
kets still rely on dedicated secure data transmis-
sion networks (Brousseau and Qulin, 1991 and
1996).

Internet Innovation

The indisputable innovation that the Internet
brings to e-commerce lies mainly in the Internet
Protocol technologies (TCP/IP, invented in the
early nineteen-seventies) and the World Wide
Web (WWW, invented in the early nine-
teen-nineties).
IP is a communications protocol that provides
interoperability for networks operating on differ-
ent standards. It makes all of the available phys-
ical networks in the world into a single network.
Network administration is also decentralised. It
provided by all of the network components that
use the IP address information included in each
packet of information travelling over the net-
work. Decentralisation means that the network
cannot be blocked and it enables the network to
grow with minimal coordination, which is lim-
ited to compliance with the Internet technical
standards. The range of functions available can
be expanded simply by hooking up more hard-
ware or introducing new software to the Inter-
net. Decentralisation also makes differentiated
communications management possible, ranging
from mass communication to point-to-point
communications and all combinations in
between in which information is targeted and
adapted to different audiences.
World Wide Web technology provides a stand-
ardised system for addressing and displaying
information that also works on all networks.
This technology can handle data, text, still
images and animated images.
We could cite hundreds of other innovations
related to the Internet, such as signal compres-
sion and encryption techniques that enhance
network capacity and security, but IP and the
World Wide Web are the two that brought about
the greatest transformation.
The Internet has become an inexpensive world-
wide multimedia network used for:
Information exchanges between economic
players in different categories, including pro-
ducers, intermediaries and consumers. These
exchanges have the potential to reorganise mar-
ket transactions,
Mass worldwide distribution of information
products and services,
Cooperation, especially between individuals
involved in creative and innovative processes
that require powerful media for informal com-
munication.
R
6

Public Networks, Corporate Networks

There are many other media for e-commerce
transactions beside the Internet. The general
public is likely to use digital television networks
and cell telephone networks for many e-com-
merce transactions in the future. Cell telephone
networks are complementary to the fixed tele-
phone networks carrying the Internet today.
They reach more people and different segments

Box 1

E-COMMERCE CATEGORIES

It is common practice to distinguish between business
and retail e-commerce applications.
B2C stands for Business-to-Consumers and refers
to sales of goods and services to the general public,
some of which are delivered online. Amazon and Dell
are the archetypes of such online retailers. However, a
large proportion of Minitel services also follow the
same pattern. B2C services are mostly provided by
commercial intermediaries, who may be new retailers,
like Amazon, or conventional retailers, like Carrefour
(Brousseau, 1999).
However, there are a few sites where manufacturers
sell their products directly to customers. There are not
many such sites, since manufacturers and service pro-
viders do not always have the necessary expertise to
sell their products and manage their customer relation-
ships. Furthermore, they do not necessarily wish to
compete with the retail outlets that handle most of their
sales.
B2B stands for Business-to-Business and refers to
transactions between businesses This category of
e-commerce is either based on bilateral agreements
between commercial partners that decide to share
information via electronic media, or else on services
provided by specialised firms. As a general rule, these
services are aimed at a particular sector or industry, in
view of the specific coordination requirements in each
industrial segment, in each country or each economic
area. These service providers are mostly offshoots of
firms in the sectors where they do business. They are
either subsidiaries of the major firms in the sector, as
is the case with the airline sector; joint-ventures com-
bining different firms, as in the case of car insurance in
France, etc.
Sometime specialised providers of

value-added

information are commissioned by industry associa-
tions or standardisation bodies to develop customised
services. For example, in the United States, GEIS, a
General Electric subsidiary developed services to pro-
mote trading between oil companies at the end of the
nineteen-seventies. B2B services are less familiar to
the general public than B2C services are, but they are
much older and involve much greater business vol-
ume.
Other

classes

of service emerged with the Internet:
C2C stands for Consumers-to-Consumers and
refers to systems aimed at promoting transactions
between individuals. There are services that operate in
the same way as classified advertisements. The inter-
active possibilities of the Internet have given rise to
genuine electronic auction markets. E-bay is the prime
example of this type of service. C2C now refers to all
of the services intermediating between individuals,
such as selling used goods, barter systems similar to
local trading systems and agencies that bring inter-
ested parties together. Such services had already
emerged with the Minitel.
Insofar as many of these services do not give rise to
commercial activity per se, it would not be appropriate
to put them all under the heading of e-commerce. Yet
the example of Napster, which a system that enables
users of the Napster software to share data stored on
all of the users hard disks, shows that such systems
can have a significant impact on commercial activities.
C2B stands for Consumers-to-Business and repre-
sents an attempt to turn the rationale governing the
relationship between supply and demand around. The
basic principle is to use electronic networks to consol-
idate individual consumers' demands and have suppli-
ers bid for their business. Priceline.com, for example,
enables individuals to announce the price that they are
willing to pay for an airline ticket. It is then up to the air-
lines whether or not to accept the prices bid. On
another level, several services are aimed at aggregat-
ing individual demands so that individual buyers can
form groups to obtain wholesale prices from manufac-
turers.
In what is bound to be a more interesting development,
there are now sites where consumers pool their infor-
mation about service providers in order to reduce the
asymmetry of information. One of the most dynamic
developments of this type is taking place in the health
information services sector in the United States. Some
sites are aimed at providing a forum for people suffer-
ing from the same diseases. In addition to their psy-
chological aspects, such

virtual communities

now
play a major economic role by promoting systems for
evaluating the quality of care, services provided by
insurance companies, treatment costs, etc. (Naba-
rette, 20021).
Some of the classification matrixes now have an extra
line and column labelled G for Government. This gives
G2B, B2G, etc. Governments have made a major effort
in recent years to provide a growing supply of informa-
tion over the Internet for economic agents (G2B and
G2C. However, such tools as the Minitel, Videotext
and EDI (electronic data interchange) had already been
used for conducting such operations in the past, albeit
on a much smaller scale. Furthermore, the relation-
ships between the government and its partners are not
commercial.
7

of the population. They can also be used when
travelling. Digital television networks can be
used for two-way communication and viewers
can order programmes through their television
sets. Cell telephone networks will offer more
services as future generations of telephones
offer higher transmission speeds. Certain seg-
ments of the population that have already
adopted earlier networks, such as the Minitel,
will continue to use them if they have no com-
pelling reason to switch to the Internet. These
media should be seen as complements to the
Internet, and not substitutes. A visit to a website
may be followed by a telephone order placed
with a call centre, a shipment to a physical
delivery point, online provision of after-sales
service, etc.
The current lack of reliability with regard to
transmission speeds and security means that the
Internet is not suited for certain types of B2B
transactions. Primarily for this reason, firms
will continue to use the dedicated data links and
networks that have been in place since the early
nineteen-seventies. In most cases, the data
transmission capacity is provided by telecom-
munications operators and the owners of physi-
cal networks, such as the French railways oper-
ator, SNCF. These networks link firms that have
entered into bilateral agreements or subscribed
to the same dedicated networks.
Once again, firms make simultaneous use of
several communications media, even though the
Internet will eventually become a platform for
integrating all network services. This future
platform will not, however, be the Internet we
know today. IP will be the common standard for
networks administered in different ways. As is
the case today, differentiated communication
services will be available. But, unlike today,
they will use a common network infrastructure,
instead of being implemented on separate phys-
ical infrastructures.
It should also be noted that systems being devel-
oped on the Internet today often rely on systems
developed a long time ago for business use.
Online financial market transaction systems and
online airline ticket sales systems are in fact
connected to the on-line trading systems and
airline reservation systems that were developed
in the mid-nineteen-seventies. Retail transac-
tion systems would not exist without these older
systems, which link major institutions and cor-
porations and ensure secure and reliable trans-
actions.
To measure e-commerce, therefore, we have to
observe what is happening on other digital net-
works, far beyond the Internet. This observation
requires us to analyse how the various channels
for exchanging information are complements or
substitutes for one another. The other channels
may be digital networks, other means of tele-
communication, the mail or face-to-face meet-
ings. These various media for information
exchanges are often thought of as substitutes for
one another, when they are in fact largely com-
plementary

6

.

Breaking Transactions Down
into their Components

t helps to analyse the uses made of electronic
networks in order to move beyond the
medium-based approach. Uses can be captured
by identifying the transactional functions that
the parties choose to carry out online, with a dis-
tinction between

ex ante

functions and

ex post

functions (Williamson, 1985).
Before a transaction can take place, the parties
must:
- find potential partners [1];
- negotiate with their partners [2];
- and then reach an agreement [3].
After the transaction has been concluded, the
agreement needs to be carried out, which
involves an exchange of information:
- specifying the parties obligations in detail
(especially if the contract is incomplete with
regard to such information as volumes, delivery
dates and addresses and certain technical speci-
fications) [4];
- enabling the parties to make adjustments
(delivery status, resource planning, etc.) [5];
- enabling the parties to reach a financial settle-
ment (invoicing and payment) [6].

6

Analysis of telephone and e-mail trafc patterns, for example,
show that the electronic exchanges are most intense between
individuals who communicate regularly by other means. In com-
mercial relationships, electronic media may also be complements
to face-to-face meetings. In the American automotive market, for
example, it is estimated that fewer than 1% of the sales are made
online, yet nearly two-thirds of the purchases are prepared on the
Internet, where buyers go for information about models and
dealers.
I
8

Finding, Negotiating and Reaching
Bargains on Digital Networks

Electronic markets emerge when the

ex ante

functions are carried out on digital networks
and supply and demand actually meet online.
Under circumstance, the parties informations
costs might be lower, which could enable them
them to make prospective counterparties com-
pete for their business. In practice, however,
contrasting forms of markets may emerge
(Brousseau, 2000). Intermediaries can organise
non-transparent markets, such as the airline res-
ervation systems, where supply and demand
information is centralised and managed accord-
ing to the intermediaries own rationales for
adjusting the other agents plans. Or, con-
versely, the market can operate as a dou-
ble-blind auction, where all bidders and sellers
are competing against each other (marketplace
rationale; see Box 2).
In the same way, the actual operating proce-
dures on the market can differ greatly, depend-
ing on whether all functions, or only certain
functions, are carried out electronically. In the
vast majority of cases, Internet-based e-com-
merce today relies mainly on an electronic
search for potential partners (Function [1]). On
the one side, product and service providers use
websites to advertise and their websites can be
seen as showcases. On the other side, consum-
ers use the capabilities of more or less special-
ised search engines that list providers and, more
rarely, compare the advertised prices. In prac-
tical terms, this limits the usefulness of such
search engines to standardised products and ser-
vices, where there is little uncertainty about
quality. Typical examples are books, compact
discs and software

7

.
The leading example in this field is the range of
service provided by the automotive sales inter-
mediaries that have emerged in the United
States (Autobytel.com, Autoweb.com, Car-
point.com). These sites list the affiliated car
dealers prices for a given model and send them
prospective buyers. They earn money by charg-
ing car dealers a subscription fee or a finders
fee. Buyers enjoy the lowest prices, which are
made possible both by keener price competition
and the economies of scale that affiliated deal-
ers achieve as a result of higher stock turnover.
A recent study (Scott Morton,

et al

., 2000)
reports that buyers save about 2% on the price of
a car.
Parties are less likely to negotiate terms and
place orders online (functions [2] [3]). Buyers
do not have much assurance as to the quality of
the products and services provided. Thus
online negotiations and orders only involve
highly standardised and not very complex
products and services, since the relevant char-
acteristics have to be described using an inter-
face that does not leave a lot of room for detail.
In addition, the products and services have to
be ones that are used in a standardised manner,
since online shoppers can only get fairly lim-
ited help in choosing the best product for their
needs

8

.
In practice, even though consultants estimates
need to be taken with a grain of salt, more than
80% of the transactions on the Internet involve
products with such clear-cut standard character-
istics. They are travel tickets, which are defined
by a departure date and city, an arrival date and
city and the means of transport, or else cultural
products, which are defined by the title and the
artist or author, and computer products, which
are defined by technical standards, capacities
and, where appropriate, generation (see
Table 3). There is a fourth category of such
products, which is financial products. A grow-
ing share of retail trading orders is being placed
via online brokerage systems.

7



8

Deliveries of these products and services do not
raise any major logistical problems. They can be
delivered electronically or via conventional cir-
cuits, such as the mail or parcel delivery serv-
ices, since such products are compact, relatively
light and do not require special treatment during
shipment and storage.

7

The main sites in America are Dealtime, mySimon, Pricewatch,
Pricescan, Pricegrabber, BottomDollar and Qixo. However, as
Ellison and Fisher Ellison (2000) point out, consumer surveys in
America shown that most of the sales made over the Internet do
not involve the use of such sites. For example, in 1997 and 1998,
70% of the buyers purchased recordings and books from a single
site (Johnson et al., 2000).

8

Several attempts to create online B2B systems failed in the
nineteen-eighties, or else they had to be redesigned. This was
because complex products and services are not easy to describe
online. For example a system for exchanging real estate market
information online was not successful since posting descriptions
and photos could not replace visits to the properties in person to
evaluate their true characteristics. In a similar vein, DEC, which
was the leader for networked mini-computers at the time, had to
limit its Electronic Store to taking orders for supplies and periph-
erals. Selling computers required DEC to give a great deal of
advice and work with customers to determine their exact require-
ments. Customers were not necessarily able to express their
needs in terms of hardware and software because they did not
have a comprehensive grasp of what was available, technical
developments and how certain solutions could match their
requirements.
9

Box 2

E-COMMERCE TECHNOLOGY

E-commerce can be based on several types of tech-
nology, which correspond in part to technological
progress. If fact, any use of an electronic medium to
carry out all or part of a commercial transaction can be
seen as an e-commerce operation. Consequently, a
fax used for placing orders can be seen as an e-com-
merce technology. A number of digital network serv-
ices are more specifically designed as media for com-
mercial transactions. Some of the main ones are:
Value added services (VAS), typified by airline reser-
vation systems and financial information and trading
systems, like Reuters. These systems operate on the
principle of a centralised server. The

clients

, which
are travel agencies and fund managers in this case,
have access to a service that is usually made up of an
information component to help them carry out their
transactions and a transaction component to record
and validate orders and to handle payments. These are
closed systems, since only subscribers have access to
them and the services offered are specific to the net-
work. Most of the time, the service provider plays an
active role, matching supply and demand, defining
precise operating rules for the market, etc. Most of the
systems used today emerged in the nineteen-seven-
ties. They were originally modelled on large centralised
mainframe systems, which were interconnected in
some cases (Brousseau and Qulin, 1991 and 1996).
Videotext, typified by France's Minitel and systems
developed in other countries in the early nine-
teen-eighties, is also based on a centralised computer
system, which was the prevailing model before the
advent of personal computers. The main differences
between videotext and value added services are that
videotext solutions are aimed at the general public and
they are much cheaper, since they use public switched
telephone networks and data transmission networks
that are accessible to everyone. Dumb terminals are
used to connect to the servers providing services. The
success of France's Minitel system is largely attributa-
ble to the free distribution of the terminals and the
invention of kiosk billing whereby the network operator
bills customers for

minutes

on the network and then
passes on a share to the information service providers.
This type of solution provides a reliable and inexpen-
sive system for handling

micropayments

that is well
suited for the provision of information services aimed
at the general public (Flichy et al., 1991; OECD, 1998).
Electronic data interchange (EDI) is typified in
France by the GALIA information sharing system used
by car makers and their suppliers, or the GENCOD
system used by major retailers. EDI is primarily based
on standardised presentation of the information
exchanges between firms so that the order processing
systems can automatically generate and interpret
messages. More importantly, EDI is linked to applica-
tions for inventory management, shipping, invoicing
and payments, as well as production and sales man-
agement. The main attraction of EDI is not automation
of information processing, but automation of coordina-
tion between firms. This requires more than just stand-
ardised message formats and involves rules governing
interfirm transactions, contracts, etc. This explains
why most EDI systems are specific to given sectors of
industry. EDI systems were developed in the late nine-
teen-eighties and, in most cases, they were integrated
into dedicated data transmission networks that were
well suited to very secure mainframe systems (Brous-
seau, 1994). However, since EDI standardisation is not
very dependent on the infrastructures used to transmit
data, WebEDI is starting to make its appearance, and
EDI messages can now be transmitted over the Inter-
net. This development will bring EDI within reach for
smaller firms since the Internet is cheaper and it can be
used to link personal computers.
Secure web servers are generally used for e-com-
merce applications. These are Internet servers that use
software techniques to protect themselves against
intrusions by hackers and to secure the information
exchanged with the computers connecting to the
server. Data security is usually provided by encryption.
In most cases, the servers currently in use offer access
to online catalogues and ordering systems that are
similar in spirit to VAS services and videotext. The
main difference lies in the Internet technology used to
connect computers to each other, instead of connect-
ing terminals to central servers. The Internet makes
multimedia presentations possible and customisation
of information so that servers send data suited to the
customers' requests or adapted to user profiles estab-
lished by the

cookies

that servers leave and read on
users' hard disks.
Servers can also use the processing and storage
capabilities of client machines for online delivery of
some types of information services (downloading).
Servers can also organise dialogues between clients'
computers. The rich potential and flexibility of web
servers mean that they are used for services operating
in very different ways (see Box 1). The first commercial
web servers, such as Amazon, were developed in the
second half of the nineteen-nineties. The fact
that services are not dependent on the medium
means that web techniques can be used to develop
applications using mobile terminals, digital television
sets, etc.
Marketplaces organise genuine virtual markets by
combining the technology of EDI with web servers and
extranet, which is the capability of creating virtual con-
trolled-access networks on the Internet. Subscribers
have secure and anonymous access to the market-
place to make business propositions to all of the
potential customers and suppliers connected to the
network. Responses are received and selected, either
by the party making the proposition or by a centralised
system for managing anonymous bids. At present,
most of the systems are still in the development or
testing phases. There are formidable technical prob-
lems to be overcome. Such systems require an
immense effort to standardise services and to come
up with rules for interfirm coordination. Partnerships
have been formed in practically every sector of indus-
try since 1998, but it is not clear that such market-
places are well suited to the specific characteristics of
transactions in every industry.
10

Obstacles to Fully Electronic Transactions

There are several obstacles to performing func-
tions [1], [2] and [3] electronically. These obsta-
cles are the reason that only a minority of trans-
actions involving a minority of products and
services are conducted as wholly electronic
transactions at this time (see Table 3).
First, in spite of technical progress, electronic
media still face the problem of coming up with
standardised language for describing products
and services (

measurement

systems in the cate-
gories proposed by North, 1990). Without such
language, consumers cannot evaluate the nature
and quality of the products and services on offer
and they cannot express their requirements. Of
course, similar problems arise in the distribution
of products and services through conventional
bricks-and-mortar networks. However, personal
contact and the physical presence of the prod-
ucts or service provider reduce uncertainty
about quality. The local presence of conven-
tional distributors and service providers also
ensures that there are means of dealing with
problems. A local operators reputation is one of
the main assurances of the quality of its service,
since its reputation serves as a form of

certifica-
tion

for the products that it distributes.
Secondly, there is also the problem of the reli-
ability of e-commerce operators (

enforcement

problems in the categories proposed by North,
1990). If they have no local presence, or are
located in a different jurisdiction than their cus-
tomers, they may be tempted by opportunism

9

as regards the quality or even the reality of what
they are offering. This can work both ways,
since customers could also escape paying for
goods and services rendered.
Thirdly, producers and consumers alike may fear
that IT will make markets too transparent and
that they will end up as the victims of this exces-
sive transparency

10

If we look at what has hap-
pened to airlines, we see that the growth of the
market for online airline reservations has intensi-
fied price competition between airlines and made
differentiation strategies pointless, with the
result that airlines are engaged in a brutal price
war (Dang Nguyen, 1996). In other industries,
operators have responded by developing strate-
gies to prevent such markets from emerging. For
example, at the end of the nineteen-eighties, the
major American insurance companies opposed
the development of on online market for insur-
ance products (Brousseau, 1993).
In addition, many producers impose restrictions
on their distributors. These restrictions are moti-

9

In transactions cost theory, opportunism means not respecting
the letter or the spirit of an agreement by misrepresenting the
quality of products or services, or else by defaulting on ones obli-
gations.

10

This argument relates more to conducting transactions on
transparent electronic markets than to conducting transactions
online per se, since sellers can set up their systems to prevent
consumers from comparing offers by differentiating offers, using
incompatible standards, setting xed subscription fees, etc.
* According to BIPE (2000), these amounts account for 7% of the book sales, 6% of CD sales and 3.7% of box office and game sales in
the United States. In the travel industry, the FRF 620 million in online sales represent 0.3% of travel and hotel sales in France.
Sources: Benchmark Group, 2000 for France (survey of the

75 most active commercial sites in France,

http://www.internet.gouv.fr/
francais / Forrester Research, 2000), for the United States, http://www.forrester.com/Home/0,3257,1,FF.html

Table 3

E-Commerce Structure in France and the United States in 1999*

Sector
France USA
FRF millions % of total USD millions % of total
Travel, transportation and hotels 620 47.18 7,798 38.33
Computers (hardware and software) 312 23.74 4,455 21.89
Cultural products (books - music - video) 137 10.43 2,376 11.67
Apparel 5 0.38 1,620 7.96
Flowers and gifts 15 1.14 656 3.23
Food and drink 33 2.51 513 2.52
Health and beauty - - 509 2.50
Furniture and appliances 25 1.90 446 2.19
Box ofce 7 0.53 300 1.47
Games 2 0.15 253 1.24
General retail (mail order, distributors, chain stores) 92 7.00 - -
Miscellaneous 66 5.02 1,418 6.97

Total 1 314 100.00 20,344 100.00
11

vated by strategic considerations aimed at creat-
ing and using market power or by the require-
ments of providing end consumers with related
after-sales service and support (OECD, 1994).
Service providers and manufacturers today
often refuse to develop distribution circuits that
would compete with their conventional
circuits

11

.
Consumers, on the other hand, may worry that
private information revealed by their online
buying habits (see below) may eventually be
used for excessively discriminatory marketing.
Many web users have already adopted the pol-
icy of giving false information when signing up
for services that are offered to users for free if
they will disclose personal information. Profes-
sionals estimate that 40% of the information
obtained in this way is false.
Beyond the shipping issues that are frequently
cited as arguments in favour of developing
mixed systems combining online and conven-
tional distribution channels (clicks-and-mortar),
information and strategic issues are cited to
explain why electronic distribution channels
(click-and-deliver) have not taken the place of
conventional channels (bricks-and-mortar)

12

. It
would be misguided to think of these elements
as just so many obstacles that need to be
removed in order to create markets that are truly
efficient because they are truly electronic. Obvi-
ously, the best media and organisational struc-
tures for commercial transactions differ,
depending on the specific characteristics of var-
ious goods and services. The most efficient
exchange procedures in the real world are not
the same for everything, as can be seen in the
variety of structures and technologies used for
offline transactions. There is no reason to think
that a single model for e-commerce can be
imposed, or even to think that e-commerce will
be suited to all types of commercial transac-
tions.

Commercial Functions of Digital Networks

Consequently, it is more important to see how
digital networks change supply-and-demand
relationships than to monitor the growth of
online ordering. Furthermore, these sup-
ply-and-demand relationships rely on several
media.
Digital networks, and the Internet in particular,
provide several useful functions for commerce.
The client-server structure of the Internet makes
it a good medium for providing personalised
information. It is a multimedia network that can
be used for carrying media content. It is decen-
tralised, which means it can facilitate the consti-
tution of self-managed communities. The com-
bination of these elements:

11



12

makes the Internet a channel for supplying
dense information that is tailored to the charac-
teristics of each individual user. In addition to
the market for specialised information (served
by the media, databanks, etc.), the Internet is a
vector for targeted marketing messages. It is
also a key to providing other services aimed
either at simplifying the tasks of potential cus-
tomers by helping them choose, make appoint-
ments, locate retail outlets, etc. or aimed at
attracting new customers by providing free
advice, for example.
makes the Internet a targeted marketing
resource. Digital network can automatically and
accurately trace and store exchanges. This gives
marketing experts the tools to draw up highly
detailed individual profiles and identify target
population groups for focused marketing cam-
paigns. This trend emerged in the nine-
teen-eighties, when computer use became wide-
spread in the retail trade. The development of
digital networks accentuated the trend by mak-
ing it possible to combine earlier databases with
monitoring of cyber-consumers. The resulting
marketing data has been used for discrimination
strategies based on identifying consumption
profiles and implementing customer loyalty
schemes (Gensollen, 2001).
makes the Internet the home for virtual com-
munities based on cooperation and information
sharing by individuals with common character-
istics. The Internet was developed in a
non-commercial context for sharing informa-
tion and knowledge. It became the medium for
the development of a number of tools, such as
e-mail, discussion lists, newsgroups and chat

11

This is especially true in some sectors, such as chain stores in
France. The retailers market power means that manufacturers
considering direct online sales have very explicit fears about
potential retaliation by retailers. In practice, manufacturers and
wholesalers e-commerce services are differentiated from their
conventional business so as to avoid competition between distri-
bution channels. For example, Compaqs online offer of comput-
ers does not include its high-end products aimed primarily at the
business market and wine dealers market imported and varietal
wines online, but they sell their highest quality products only
through conventional distribution channels (Giraud-Hreaud et
al., 2001).

12

In the early days of e-commerce, the contrast was between the
click-and-deliver and the bricks-and-mortar models. Thus the
model combining conventional distribution channels with the
capabilities of digital networks was dubbed clicks-and-mortar.
12

rooms. These tools have come to play a key role
in managing relationships between supply and
demand. In the past, consumers organised them-
selves into groups in an attempt to reduce the
asymmetry of information that put them at a dis-
advantage (see C2C in Box 1). But sellers are
now using the power of these tools for their own
ends by providing such tools to their customers
and building communities around their products
and services (Curien

et al

., 2001). Amazon.com
is a prime example, since it asks its customers to
review and rate books to help guide other con-
sumers in their choices. Software vendors use
newsgroups to help identify bugs in products,
thus making each customer a product tester.
They also provide user support newsgroups that
ease the workload of their after-sales service
departments. These tools also help suppliers to
identify the product features that consumers
want.
If we measure only agreements entered into
online (function [3]), then we only account for a
fraction of the applications having an impact on
the way markets work and we run the risk of
underestimating the proportion of operations to
find partners (function [1]) and negotiate deals
(function [2]) that are carried out online.
On the other hand, it does not make sense to
simply aggregate all of the sales volume that
involves digital networks in one way or another,
without making distinctions between the opera-
tions concerned. In the first place, such an
aggregation would not enable us to determine
which aspects of transactions are affected. Fur-
thermore, the use of digital media would soon
cease to be a discriminating variable, since such
media will be used as the medium for one func-
tion or another in nearly 100% of the cases, as is
the case with the telephone network today.
Relationship Management
over Digital Networks
After an agreement has been entered into, elec-
tronic media become powerful tools for manag-
ing logistical coordination between the con-
tracting parties (function [4]) and enabling the
parties to provide each other with information
services(function [5]).
One of the advantages of digital networks for
coordination as part of function [4] is that they
make automated information exchanges possi-
ble. They can link automatic systems for man-
aging interfaces between economic partners.
This is done to manage purchasing and sales,
inventories and shipping, and even production
scheduling. This helps to reduce relative ex post
adjustment costs and potential response times
13
.
The most striking example in this area has been
the development of EDI in the automotive
industry. The need to combine mass production
and greater product differentiation led to the
implementation of flexible just-in-time produc-
tion methods. As more than 70% of the value
added in a car comes from the carmakers sup-
pliers, it became critical to make inter-firm rela-
tionships more flexible and fluid. The old sup-
ply contracts, which stipulated the technical
specifications and delivery schedules for prod-
ucts in the medium term, have been replaced by
less complete contracts, where delivery types
and schedules are redefined on an ongoing
basis, according to the carmakers sales and pro-
duction requirements. These new production
and coordination techniques would probably
not be feasible without the development of the
EDI systems that provide very close links
between the suppliers and the carmakers com-
puterized production management systems
(Brousseau, 1994). The same type of
just-in-time sourcing is now the prevailing prac-
tice in retail distribution.
Electronic media are also effective tools for pro-
viding personalised information services (func-
tion [5]). These media are either the main vector
for the services on offer, as is the case with
online information services, or else they are the
vector for differentiating and enhancing serv-
ices, as it the case of services developed by
transportation firms. Many express delivery
firms now offer customers real-time tracking of
their shipments, so that customers can deploy or
re-deploy their resources preventatively.
Most of these information services are sold on a
subscription basis (which may be combined
with per-use charges). The idea is to lock in cus-
tomers by providing a service for which there is
no substitute or by bundling several comple-
mentary information services into a subscrip-
tion service. Bundling actually presents many
advantages that are well know in terms of indus-
13
Technical potential must not be confused with organisational
capability. A great deal of research on the use of ITC has shown
that rms have to reorganise to achieve the full potential of ICT in
terms of exibility. In this particular case, the reorganisation con-
cerns inter-rm interfaces. Organisational changes are a delicate
matter and very costly (Brousseau, 1994; Brousseau and Rallet,
1999).
13
trial economics, particularly with regard to dis-
criminating customers and building customer
loyalty (Gensollen, 1999; Gaudeul and Julien,
2001). Bundling information services makes it
possible to solve security problems in informa-
tion transactions, since the products that are less
vulnerable to resale ensure the sale of more vul-
nerable products. Bundling also makes it possi-
ble to implement marketing strategies based on
the tracking of consumer profiles mentioned
above.
This is the Dawn of Alternative
Payment Systems
Making payments via electronic media [6] is
fast and cheap. Implementation problems are
primarily related to organisational and institu-
tional issues. In addition to access to secure net-
works and reliable encryption systems, the main
problems lie in the development of sufficiently
all-purpose systems so that the payment data
exchanges are really universal equivalents and
are reliable enough to maintain users confi-
dence. In other words, the new technical tools
and, where appropriate, new service providers,
have to be fitted into the conventional means of
payment management system run by the banks
and overseen by the banking authorities.
The emerging alternative payment systems,
such as e-cash, secure protocols, private clear-
ing systems, notarised systems and loyalty
reward systems, all raise different issues, since
they have differing impacts on monetary crea-
tion, bypassing the banking system, etc. Such
systems are still in their infancy, which is why
so few online payments are made over open net-
works. The use of conventional means of pay-
ment, and particularly the practice of giving
credit card numbers online, still raises major
security problems.
On the other hand, payments over secure net-
works are already widely developed. A growing
share of such payments is made on customers
instructions to their banks. Interbank networks
developed in the nineteen-seventies, including
the international SWIFT network, are widely
used for transfers between bank accounts. The
main challenge today lies in the development of
online payment systems that can reliably handle
large numbers of small-value payments for the
general public (Bounie, 2001). This may entail
the development of genuine private cross-bor-
der monetary-issuance systems. Consequently,
online payment solutions raise very sensitive
issues
14
.
All in all, the problems involved in the use of
digital networks for ex post transactions man-
agement are very different from those involved
in ex ante transactions management.
Ex post transactions management systems
have nothing to do with finding partners or
negotiating bargains. Consequently, they have
no direct influence on the way markets operate,
which means their only impact should be on the
efficiency of coordination. Nevertheless, such
systems are so carefully tailored to fit the par-
ties specific needs very closely that they lock in
the partners in the transactions concerned.
Groups of users sharing the same technical and
organisational standards (Brousseau, 1994)
become de facto closed communities. Further-
more, the inter-organization information shar-
ing systems provide services that are designed
for differentiation, thereby increasing switching
costs for users.
Wholesale payments are already handled
electronically on the whole, as are most transac-
tions between banks and their business custom-
ers. This means that the critical issue now with
regard to the development of electronic pay-
ment systems for emerging markets, such as
online information providers, is the design of
institutional frameworks for regulating pay-
ments between consumers and businesses.
Once again, the criterion of online ordering is a
poor choice for capturing the influence of IT on
coordination procedures in market economies.
Categories of Players and Differentiation
of E-Commerce Forms
Another way of approaching IT uses is to iden-
tify the types of players that the different elec-
tronic commerce services bring together (see
Box 1). The most common distinction between
B2B and B2C relates to the transaction volume
handled or likely to be handled via electronic
systems. All of the measurements and forecasts
estimate that online B2B business volume (see
Table 4) is 8 to 10 times greater than online B2C
14
Online securities trading systems raise similar problems.
Online brokers are capable of handling clearing amongst them-
selves and are setting up systems that bypass the conventional
capital markets. This raises regulatory issues with regard to secu-
rities markets, even though competition between the various
markets creates greater efciency (Serval, 2001).
14
business volume. However, examining e-com-
merce in these terms raises problems (see
above), especially as B2B transactions are very
different from B2C transactions at present.
Most observers agree that B2C primarily con-
cerns the operations leading up to a contract
(functions [1], [2] and [3]). Most of the existing
systems were designed to handle search and
negotiation operations relating to a very narrow
range of goods and a tiny proportion of transac-
tions (see Table 3). Nonetheless, B2C sites play
a substantial role in the markets concerned. The
spectacular failure of some e-commerce busi-
nesses launched in other markets, such as cloth-
ing, suggest that the first B2C market segments
were not so much trailblazers as segments that
lent themselves particularly well to online han-
dling of some or all of the operations leading up
to contracts (see above).
The banking and financial sector offers the most
substantial range of ex post transactions man-
agement in the B2C market. Banks first intro-
duced home banking systems back in the
mid-nineteen-eighties, when American house-
holds were buying personal computers and
French households gained access to telematics
through the Minitel system. The advent of the
Internet as a commercial network greatly
increased the capacities of such systems and
opened the field to new operators, who brought
sweeping changes in the way households trade
on capital markets. Online trading systems
make it possible for members of the general
public to manage their market positions in
real time. Online securities trading now
accounts for some 45% of trading in the United
States, leading to a fall in brokerage com-
missions of some 70%, according to the BIPE,
as a result of greater competition and the sav-
ings on back office
15
costs in online trading sys-
tems.
While B2C e-commerce mainly focuses on ex
ante transactions management, B2B focuses
more on ex post transactions management. The
1999 survey of French manufacturing firms by
the Census Department of the French Ministry
of Industry (Sessi) shows that online ordering is
tiny in comparison to information exchanges
between industrial and commercial partners
(see Table 5 and Chart). The exceptions are the
financial and airline sectors, which developed
B2B online transaction systems in the nine-
teen-seventies. The oil industry also developed
online brokerage systems in the wake of two oil
shocks to manage a good that had come to have
a very high unit price.
15
The vast majority of existing systems, such as
EDI systems, were designed to handle ex post
transactions management, including order
scheduling, sharing technical information, syn-
chronisation of production and shipping and
handling invoicing and payments. As in the case
of B2C, e-commerce transactions primarily
concerned specific activity sectors. For exam-
ple, Edifrance estimates that two sectors
account for most of the FRF 800 billion volume
of commercial EDI traffic in France: with distri-
bution accounting for FRF 500 billion and the
automotive sector accounting for FRF 200 bil-
lion (Observatoire du Commerce et des
changes lectroniques, 1999)
16
.
But these two sectors are not the only ones to
use e-commerce. In addition to EDI, there are
many other systems for exchanging information
between business that are based on less highly
automated means of communication, such as
the Minitel, or on less highly organised means
of communication, such as e-mail. These sys-
tems can play a role in transactions management
in other sectors. For example the Minitel plays a
role in the transport sector (Brousseau, 1991).
There are also long-established online informa-
tion systems that rely on the dedicated networks
that information service providers use to deliver
their services to business customers (see
Table 6).
15
The nancial sector makes a distinction between front ofce
operations, which are necessary for carrying out trades on the
markets, and back ofce operations, which are the internal and
administrative procedures for preparing and processing trades.
16
It should be noted that EDI projects abounded in all sectors at
the end of the nineteen-eighties, but EDI only plays a major role
in two very specic areas today. The pattern is likely to be similar
for Internet technology. The possibilities opened up by various
forms of IT and by the various potential applications are not nec-
essarily useful or applicable in all areas. This means there is little
reason to believe that a single model of e-commerce will become
the norm in all areas.
* The estimates are based on a definition of e-commerce where
online operations are any transactions carried out over a data
transmission network (Internet, EDI, etc.), regardless of the pay-
ment method used.
Source: Forrester Global eCommerce Model, 2000.
Table 4
Online Business Volume in 2000*
In USD billions
USA Japan France
B2B 449.900 29.618 9.102
B2C 2.262 38.755 818
Total 488.655 31.880 9.920
15
The Effects of E-Commerce
Take Many Shapes, Which
are Poorly Understood
ar from bearing out the myth of the Internet
as a marketplace, the emerging pattern of
e-commerce is one of marginal use of differing
systems for:
- supporting and guiding consumers seeking
suppliers and products (and ordering online
occasionally), with B2C e-commerce used for
very limited range of products;
- more efficient management of industrial and
commercial partnerships with B2B e-com-
merce;
- providing online information services.
Naturally, qualitative and quantitative devel-
opments are to be expected. These should
eventually have a major impact on the way
markets and industries work, but it will vary
from area to area, since different systems are
being developed to work in different contexts.
How will these developments affect future
markets?
F
Field: manufacturing firms with 20 or more employees, including the energy industry and excluding the agri-food industry.
Source: TIC survey, 1999, Sessi.
Chart
The Advent of E-Commerce
Key: The percentage of firms offering customers online ordering (via EDI, web sites or minitel) and online payments (via EDI, web sites or
minitel) are shown for each of the two groups (all manufacturing firms and large firms with more than 500 employees).
Field: manufacturing firms with 20 or more employees, including the energy industry and excluding the agri-food industry.
Source: TIC survey, 1999, Sessi.
Table 5
Electronic Data Interchange (EDI)
As % of rms
Number of employees
Total
20 to 49 50 to 99 100 to 249 250 to 499 > 500
With other rms (subcontractors,
suppliers, customers, banks, etc.) 29 39 47 60 76 36
Between divisions within the rm 10 16 21 31 40 15
With government agencies
and public bodies 11 14 20 26 34 14
For one of the three functions 36 47 57 69 82 44
0 3 6 9 12 15 18 21 24 27 30 33 36
As % of firms with more than 20 employees
EDI ordering
Web site
Minitel
EDI payments
Web site
Minitel
EDI ordering
Web site
Minitel
EDI payments
Web site
Minitel A
l
l

m
a
n
u
f
a
c
t
u
r
i
n
g

f
i
r
m
s


L
a
r
g
e

f
i
r
m
s
16
Two Generations of Research
Rigorous research in this area is still rare. This
is undoubtedly due the relative newness of the
phenomenon. However, there is nothing new
about using digital networks as a medium for
commercial transactions (see above). Electronic
markets and data exchanges between companies
have been around for more than 20 years. There-
fore, we need to draw a distinction between two
generations of research.
The first generation started in the nine-
teen-eighties and analysed the impact of
inter-organisation information sharing systems
(e.g., Barette and Konsisky, 1982; Brousseau,
1993; Clemons and Row, 1988; Faulhaber,
Noam and Tasley, 1986; Malone, Yates and
Benjamin, 1987; Scott Morton, 1991). This
research was mainly intended to determine
whether ICT brought about changes in the gov-
ernance of inter-organisation transactions
through disaggregation of hierarchies into
markets, changes in the economics of certain
industries or productivity and competitiveness
gains. Most of the research focused on analysis
of B2B e-commerce.
At the end of the nineteen-nineties, a second
generation of research started to analyse the
impact of the development of e-commerce on
the Internet. Very few findings have been pub-
lished, since the advent of Internet-based
e-commerce is still very recent. Most of this
research focuses on B2C e-commerce, since the
development of Internet B2B e-commerce is
still in its infancy
17
. In both cases, we do not
have enough perspective to observe the impact
of different types of e-commerce applications
on the Internet. The potential organisational
changes triggered by these applications have
barely started to be seen and, more importantly,
the services offered and the strategies used are
changing constantly. Applied research based on
proper scientific foundations is therefore rare,
which fuels a great deal of purely speculative
writing about the presumed impact of the Inter-
net and the advent of B2C e-commerce.
17
The context for the second generation of
research is different from that of the first. The
focus is primarily on testing the reality of the
myth of the Internet being the medium for com-
petitive markets with no intermediaries.
These two issues (competitive markets and dis-
intermediation) need to be clarified before tak-
ing a more general look at the type of research
that needs to be promoted.
Prices are not Always Lower
on the Internet
Some research has been done on retail sites in
the last five years to test the reality of the myth
that IT has made markets more transparent. This
type of research raises major methodological
problems, because there is no guarantee that the
products and services being compared are truly
equivalent and because web sites do not always
give prices. The research also has to focus on a
17
Even though the volume of B2B trafc is greater than B2C traf-
c, most of the B2B trafc consists of EDI exchanges relating to
ex post transactions management. Internet-based e-procure-
ment, with the development of virtual marketplaces (see Box 2),
is just starting to get off the ground. Therefore, looking at the B2C
activity gives us more perspective on how large-scale virtual mar-
kets work than observation of the B2B market does.
Source: TIC survey, 1999, Sessi.
Table 6
Manufacturing Firms and the Internet
As % of manufacturing rms
Internet access Web site Online ordering Intranet Extranet
Major rms
SMEs
97.7
67.6
70.5
38.1
13.6
8.9
79.1
19.3
37.2
8.6
High tech
Low tech
85.6
64.3
51.5
36.2
9.2
8.9
36.5
17.4
16.8
7.8
Export rate over 25%
Export rate under 25%
83.0
64.0
52.4
35.0
10.2
8.6
33.4
17.3
16.7
7.3
Group subsidiaries
of which foreign group subsidiaries
Independent rms
82.1
85.7
59.2
47.9
46.3
33.3
9.5
9.8
8.6
38.4
46.5
9.5
15.8
20.0
5.3
Innovative rms
of which innovative product rms
Non-innovative rms
80.5
81.6
58.7
49.6
51.1
30.5
11.1
11.3
7.3
28.6
29.5
15.2
13.2
13.5
6.6
17
very specific range of products, which are pri-
marily books and compact discs. Furthermore,
research has led to contrasting findings, which
make us think that reliance on electronic media
is far from being the sole factor to consider in
current developments and that other aspects,
such as market structures and marketing strate-
gies, also need to be considered.
The first study by Bailey (1998a and 1998b) was
based on 24,000 prices collected on the American
market for books, compact discs and software in
1997. This study concluded that prices are not
lower on the Internet than they are in stores. It fur-
ther found that price dispersion is greater on the
Internet and that price adjustments are more fre-
quent. According to Bailey, Internet price levels
are explained by the profile of Internet consum-
ers, who have high educational attainment and
income levels and who use the Internet not so
much to find the best bargains as to save time.
Many other observations have upheld this view
(see footnote 7). Contrary to common belief, it is
relatively expensive to find information on the
Internet
18
and most of the consumers using the
Internet at this time are more concerned with sav-
ing time than they are with saving money.
At this stage in the development of the Internet,
we tend to see distribution channels using the
Internet more to target specific types of con-
sumers rather than consumers using the Internet
to shop for the lowest prices and increase com-
petition between sellers. A great deal of subse-
quent research has borne out Baileys observa-
tions: Degeratu, Rangaswamy and Wu (1998)
on grocery prices; Shankar et al. (1998), along
with Clemons, Hann and Hitt (1998), on travel
prices and Lee (1998) on used car prices.
However, more recent and more in-depth inves-
tigations have added greater detail to the earlier
observations. In some cases (Scott Morton et al.,
2000), Internet shoppers do enjoy lower prices
(see above). Brynjolfson and Smith (1999)
looked at 8,5000 prices for books and compact
discs to conclude that prices on the Internet are
9% to 16% lower than in stores. On the other
hand, their research also showed that the price
levels are proportional to the prominence of
web sites and traffic levels. Thus, despite offer-
ing lower prices, less well-known sites are not
able to attract Internet shoppers away from the
top sites, which is proof that the market is not
perfectly transparent.
In fact, more recent research (e.g. Friberg, Gan-
slandt and Sandstrom, 2000) has shown that the
Internet is a means of implementing personal-
ised pricing policies. This observation is in line
with the more qualitative observations that the
possibilities of the Internet make mass-custom-
isation policies feasible. These policies aim for
a subtle combination of differentiation and tar-
geting (see above and Gensollen, 2001). Inter-
net shoppers enjoy services that are adapted to
their main needs, which are saving time, shop-
ping from home and obtaining expert advice. In
exchange, consumers let sellers adjust prices
according to their propensity to pay.
18

No Progress towards Perfect
Virtual Markets
Up until the bursting of the dot-com stock mar-
ket bubble, the press and some of the Internet
gurus thought that disintermediation was a
key to productivity gains stemming from
e-commerce. The thinking was that digital net-
works would squeeze conventional commercial
intermediaries out of business by making them
obsolete. Events have not borne out this widely
held belief. Far from being dominated by man-
ufacturers and service providers, e-commerce is
mainly a business for conventional intermediar-
ies, such as chain stores, specialised retailers
and mail-order firms. The Internet is obviously
a complementary channel for these retailers,
alongside their other channels for selling goods,
rendering services and finding customers.
This stands to reason when we look at the eco-
nomics of intermediation in commerce. Far
from being intermediaries that merely provide
information, commercial intermediaries pro-
vide shipping, financial services and security.
These services are closely linked (economies of
scope) and require specific know-how (Spulber,
1996). The advent of new media has modified
the economics of commercial intermediation,
but it has not resulted in radical change (Brous-
seau, 2002). Thus, for reasons that relate to both
logistics and strategy (Dang NGuyen, 1999;
Rallet, 2000; Gensollen, 2001), intermediaries
in commerce continue to enjoy a competitive
advantage in the intermediation function, as
long as they are able to adapt digital technolo-
gies and make them complementary to their
conventional tools.
18
One method for cutting search costs is to use Shopbots
(Shopping Robots) which are smart search engines that shop the
Internet for the e-commerce sites with the lowest prices. How-
ever, as Bailey (1998a and 1998b) points out, e-commerce site
administrators tend to prevent shopping bots from accessing
their sites.
18
There was never any guarantee that many of the
new intermediaries created in the wave of inno-
vative start-ups over the last five years would be
viable. Revenues failed to cover costs and the
survival of many start-ups depended on the
new-economy stock market bubble. Now that
the bubble has burst, their continued existence is
in jeopardy. The start-ups are either being taken
over by conventional intermediaries or they are
going bankrupt. Some of these firms managed
to build up world brand images, as in the case of
Amazon. They are bound to survive and become
leading retailers. Eventually, however, they are
likely to become multi-channel retailers just like
their competitors are. They will have to
acknowledge the fact that the most efficient dis-
tribution channels vary depending on popula-
tion densities and consumer habits and that
commercial and logistical systems vary from
one country to the next.
Recent research on Internet-based e-commerce
tends to discredit the assumption that we are
moving towards perfect virtual markets. E-com-
merce takes contrasting forms in different sec-
tors. It complements rather than replaces con-
ventional distribution channels and technology.
Players turn it to their own purposes. This is par-
ticularly true of commercial intermediaries,
who use e-commerce to serve their differentia-
tion and targeting strategies. These strategies
produce very different levels of profitability and
very different divisions of the surplus in various
competitive environments.
This recent stream of research upholds one of
the main conclusions of earlier research on
inter-organisation information sharing systems.
ICT per se is not a determining factor. Every-
thing depends on how ITC is used and on the
environments where it is implemented. When
inter-organisation information sharing systems
were deployed, there was a strengthening of
quasi-integration around dominant firms in
some industries and, simultaneously, disaggre-
gation of other forms of industrial organisation.
Similarly, the measurement of the impact of
e-commerce in the future is likely to depend
largely on the types of systems deployed in var-
ious environments.
Lessons for Statisticians
Debate is still continuing at the French Coucil
for Statistics (CNIS), Eurostat and the OECD
(CNIS, 2000), but the tendency is to prefer a
very restrictive definition of e-commerce, with
the criterion of online ordering, as the basis for
the statistical observations now being devel-
oped. From a technical point of view, such a
definition has the advantage of being clear and
simple. From the economic point of view, how-
ever, all of the previous discussion stresses that
it will lead to a strongly biased view of the influ-
ence of ICT on coordination procedures
between firms and between businesses and con-
sumers.
This bias could be detrimental for macroeco-
nomic and microeconomic analysis of current
changes. The issue of faster factor productivity
growth and its relationship to ICT diffusion (the
productivity paradox) is still largely an open
question, since, with a few exceptions, the
research done up until now lacks basic micro-
economic data about organisational changes
related to ICT. Even though more recent
research has tended to fill this gap with regard to
coordination within firms, we will need more
reliable data to examine changes in relation-
ships between firms and between businesses
and consumers.
We need more microeconomic data to inform
players and public authorities strategies. It
will be difficult to promote the most promising
forms of e-commerce if we do not have tools for
measuring how using digital technologies to
handle various transaction functions affects new
organisational structures in commercial inter-
mediation and thus employment and the divi-
sion of labour in the retail and shipping sector ,
as well as productivity gains, and commercial
practices with regard to differentiation and dis-
crimination strategies and thus the division of
the surplus among the various players in the
system , etc.
All of these are good reasons to use investigative
methods that go far beyond simply measuring the
business volumes affected by the use of digital
networks. Surveys of e-commerce need to come
up with better measurements of the various cate-
gories of uses and more systematic evaluation of
the business environments (particularly market
structures and industrial organisation) where
such uses are implemented, and their impact
(particularly on organisational structures). Fur-
thermore, a comprehensive set of measurements
is needed to conduct more relevant analyses of
the microeconomic and macroeconomic impact
of e-commerce. In addition to looking at price
levels and price dispersion, or transactions costs,
which are not the same as information costs, we
need to enhance the analysis of the impact of
19
e-commerce systems on the quality of distribu-
tion services, on positive and negative externali-
ties, on suppliers efficiency, etc.
These are the keys to developing an accurate
analysis of the impact that new digital tools
have on price formation, industrial structures,
competition structures, households welfare, the
efficiency of production systems, etc. This
effort is bound to require targeted surveys rather
than general statistical classifications, which are
unlikely to be accurate because of the diversity
of uses and practices from one sector to the next
and from one country to the next. We need to
adapt our methods and measurements to the
diverse range of e-commerce and the different
ways of organising transactions in different con-
texts. It would also be worthwhile to look into
the possibility of putting together reliable statis-
tical series on the development of e-commerce,
now that we have some historical perspective on
some of its forms.
The author would like to thank the Forum members and two readers, whose remarks and suggestions
helped to improve a previous version of this article.
REFERENCES
Arthur, B. W. (1989), Competing Technologies,
Increasing Returns and Lock-In by Historical
Events, Economic Journal, 99, pp. 116-131.
Bailey, J. P. (1998a), Electronic Commerce: Pri-
ces and Consumer Issues for Three Products:
Books, Compact Discs and Software, DSTI/
ICCP/Ie(98)4, OECD, Paris.
Bailey, J. P. (1998b), Intermediation and Electro-
nic Markets: Aggregation and Pricing in Internet
Commerce, Ph.D., Technology, Management and
Policy, Massachusetts Institute of Technology,
Cambridge, Ma.
Barett, S. and B. Konsisky (1982),
Inter-Organization Information Sharing Sys-
tem, MIS Quarterly, Dec., 6, pp. 93-104.
BIPE (2000), Bonjour la nouvelle conomie ! Les
enjeux prospectifs de la rvolution numrique,
BIPE, Paris.
Bounie, D. (2001), Monnaie lectronique: nature
conomique et incidences montaires Revue co-
nomique, Numro Hors-Srie, Economie de
l'Internet, sld. E. Brousseau & N. Curien, octobre
2001, vol. 52, pp. 313-330.
Brousseau, E. (1991), Telematics and Professio-
nals: from the Absence of Discrimination to the
Limitation of Uses, in Flichy, P., J. Jout and P.
Beaud (eds), European Telematics, the Emerging
Economy of Words, North-Holland, Amsterdam,
pp. 33-59.
Brousseau, E. (1993), Lconomie des contrats ;
technologies de linformation et coordination
interentreprises, PUF, Paris.
Brousseau, E. (1994), EDI and Inter-Firm Rela-
tionships, Toward a Standardization of the Coordi-
nation Processes?, Information Economics and
Policy, vol. 6, No. 3-4, pp. 319-347.
Brousseau, E. (2002), The Governance of Tran-
saction by Commercial Intermediaries: An Analy-
sis of the Re-engineering of Intermediation by
Electronic Commerce, International Journal of
the Economics of Business, 8(3), forthcoming,
Brousseau, E. (2000), What Institutions to Orga-
nize Electronic Commerce: Private Institutions
and the Organizations of Markets, Economics of
Innovation and New Technology, 9:3, July-Sep-
tember, pp. 245-273.
Brousseau, E. and B. Qulin (1991), Stratgie
de coopration et formation dun march : les nou-
veaux services de tlcommunications, in colla-
boration, Revue dconomie Industrielle, No. 56,
2
nd
quarter 1991, pp. 77-100.
Brousseau, E. and B. Qulin (1996), Asset Spe-
cificity and Organizational Arrangements: The
Case of the New Telecommunications Services
Market, Industrial and Corporate Change,
vol. 5, No. 4, pp. 1205-1230.
Brousseau, E. and A. Rallet (eds) (1999), Tech-
nologies de lInformation, Organisation et Perfor-
mances conomiques, Commissariat Gnral du
Plan, Paris, September. (http://atom2.univ-paris1.
fr/FR/membres/eric/eric.htm).
Brynjolfsson, E. and M. D. Smith (1999), Fric-
tionless Commerce? A Comparison of Internet
and Conventional Retailers, Proceedings of
20
Understanding the Digital Economy Data, Tools
and Research, U.S. Department of Commerce,
Washington, D.C., May 25-26. (http://ecom-
merce.mit.edu/papers/friction).
Clemons, E. K., I. H. Hann and L. M. Hitt
(1998), The Nature of Competition in Electronic
Markets: An Empirical Investigation of Online
Travel Agent Offerings, Working Paper, The
Wharton School of the University of Pennsylva-
nia, June.
Clemons, E. K. and M. Row (1988), McKesson
Drug Co.: a Case Study of Economost, a Strategic
Information System, Journal of Management
Information Systems, 5:1, Summer, pp. 36-50.
CNIS (2000), Rapport du groupe de travail du
CNIS (Conseil National de lInformation statisti-
que) sur lobservation statistique du dveloppe-
ment des TIC et de leur impact sur lconomie,
CNIS, Paris.
Cowan, R. (1990), Nuclear Power Reactors: a
Study in Technological Lock-In, The Journal of
Economic History, 3.
CRIC (1999), Exploring the Effects of E-Com-
merce, Report Retail and Consumer Services
Foresight Panel, ESRC Centre for Research on
Innovation and Competition, University of Man-
chester & UMIST. (http://les.man.ac.uk/cric/
e-commerce).
Curien, N. (2000), Lconomie des rseaux, La
Dcouverte, Collection Repres, Paris.
Curien, N., E. Fauchart, G. Laffond, J.
Lesourne and F. Moreau (2001), Forums de
consommation sur lInternet : un modle volu-
tionniste, Revue conomique, Numro
Hors-Srie, Economie de l'Internet, sld. E.
Brousseau & N. Curien, octobre 2001, vol 52, pp.
119-136
Dang NGuyen, G. (1996), Les systmes de
rservation arienne et lconomie des rseaux,
in Brousseau, E., P. Petit and D. Phan (eds),
(1996), Mutations des Tlcommunications et
Rorganisation des Activits et des Marchs, Eco-
nomica, Paris, pp. 231-261.
Dang NGuyen, G. (1999), Du commerce lec-
tronique lintermdiation lectronique, in
Brousseau, E. and A. Rallet (eds), Technologies de
lInformation, Organisation et Performances co-
nomiques, Commissariat Gnral du Plan, Paris.
David, P. A. (1985), Clio and the Economics of
QWERTY, American Economic Review, 75:2,
May, 332-7.
Degeratu, A., A. Rangaswamy and J. Wu
(1998), Consumer Choice Behavior in Online
and Regular Stores: The Effects of Brand Name,
Price and Other Search Attributes, Presented in
Marketing Science and the Internet, INFORM
College on Marketing and Mini-Conference.
Cambridge, Ma. 6-8 March.
Ellison, G. and S. Fisher Ellison (2000), Search,
Obfuscation and Price Elasticities on the Internet,
mimeo, MIT & NBER.
Eurostat (2001), Statistiques sur la Socit
dinformation, Statistiques en Bref, Thme 4,
4-2001, Eurostat.
Faulhaber, G., E. Noam and R Tasley (eds)
(1986), Services in Transition, the Impact of ICTs
on the Service Sector, Ballinger, Cambridge, Ma.
Flichy, P., J. Jout and P. Beaud (eds) (1991),
European Telematics, the Emerging Economy of
Words, North-Holland, Amsterdam.
Friberg, R., M. Ganslandt and M. Sandstrom
(2000), E-Commerce and Prices: Theory and
Evidence, Stockholm School of Economics, EFI.
Guaudeul, A. and B. Julien (2001), E-Com-
merce, vers une analyse conomique, Revue co-
nomique, Numro Hors-Srie, Economie de
l'Internet, sld. E. Brousseau & N. Curien, octobre
2001, vol 52, pp. 97-118
Gensollen, M. (1999), La cration de valeur sur
Internet, Rseaux, 17(97), November, pp. 15-76.
Gensollen, M. (2001), Internet : march lectro-
nique ou rseaux commerciaux ?, Revue cono-
mique, Numro Hors-Srie, Economie de l'Inter-
net, sld. E. Brousseau & N. Curien, octobre 2001,
vol 52, pp. 137-164
Giraud-Hreaud, E., L. G. Soler and H. Tanguy
(2001), Internet et la distribution des biens physi-
ques : analyse de lmergence de nouvelles struc-
tures verticales dans le secteur du vin, Numro
Hors-Srie, Economie de l'Internet, sld. E.
Brousseau & N. Curien, octobre 2001, vol 52, pp.
213-234
Johnson, E. J., W. Moe, P. Fader, S. Bellman
and J. Lohse (2000), On the Depth and Dyna-
mics of World Wide Web Shopping Behavior,
21
Columbia University Graduate School of Busi-
ness, mimeo.
Katz, M. and C. Shapiro (1985), Network
Externalities, Competition and Compatibility,
American Economic Review, June 424-40.
Lee, H. G. (1998), Do Electronic Marketplaces
Lower the Prices of Goods?, Communications of
the ACM, 41(1), pp. 73-80.
Lorentz, F. (1998), Commerce lectronique :
une nouvelle donne pour les consommateurs, les
entreprises, les citoyens et les pouvoirs publics,
Rapport du Groupe de Travail sur le Commerce
lectronique, Ministre de lconomie, des Finan-
ces et de lIndustrie, Paris.
(http://www.finances.gouv.fr/
mission_commerce_lectronique/rapports/).
Lorentz, F. (1999), Commerce lectronique :
une nouvelle donne pour les consommateurs, les
entreprises, les citoyens et les pouvoirs publics,
Rapport du Groupe de Travail sur le Commerce
lectronique, Ministre de lconomie, des Finan-
ces et de lIndustrie, Paris.
(http://www.finances.gouv.fr/
mission_commerce_lectronique).
Lucking-Reiley, D. and D. F. Spulber (2000),
Business-to-Business Electronic Commerce,
mimeo, Vanderbilt University and Northwestern
University.
Malone, T. W., J. Yates and R. Y. Benjamin
(1987), Electronic Markets and Electronic Hie-
rarchies, Communications of the ACM, 30(6),
484-497.
Nabarette, H. (2002), Internet en sant : quelle
utilisation pour les patients ?, Rseaux,forthco-
ming.
North, D. C. (1990), Institutions, Institutional
Change and Economic Performance, Cambridge
University Press.
Observatoire du Commerce et des changes
lectroniques (1999), Les Chiffres cls du Com-
merce lectronique.
(http://www.edifrance.org/).
OECD (1994), Les accords de franchise, OECD
publication.
OECD (1998), Lexprience franaise du Minitel:
leons pour le commerce lectronique, Groupe de
travail sur lconomie de linformation, DSTI/
ICCP, Paris.
OECD (2000a), Internet and Electronic Com-
merce Indicators Update,
(http://www.oecd.org/dsti/sti/it/ec/index.htm).
OECD (2000b), The Economic and Social
Impacts of Electronic Commerce: Preliminary
Findings and Research Agenda,
(http://www.oecd.org/e_commerce/sum-
mary.htm).
Rallet, A. (2000), Le commerce lectronique et
lvolution des modles de distribution et de pro-
duction, Sminaire 2000 de la Direction des Sta-
tistiques dEntreprise de lInsee, e-commerce :
Quels faits ? Quels effets ?, 6 December 2000.
Scott Morton, F. F. Zettelmeyrf and J. Risso
(2000), Internet Car Retailing, mimeo, Yale
University & NBER.
Scott Morton, M. S. (ed) (1991), The Corpora-
tion of the 1990s, Information Technology and
Organizational Transformation, Oxford Univer-
sity Press, Oxford, UK.
Serval, T. (2001), Lorsque les rseaux d'infor-
mation deviendront des bourses,, Revue cono-
mique, Numro Hors-Srie, Economie de l'Inter-
net, sld. E. Brousseau & N. Curien, octobre 2001,
vol 52, pp. 249-266
Shankar, V., A. Rangaswamy and M. Pusateri
(1998), The Impact of Internet Marketing on
Price Sensitivity and Price Competition, Paper
Presented at Marketing Science and the Internet,
INFORM College on Marketing Mini-Confer-
ence. Cambridge, Ma. 6-8 March.
Shapiro, C. and H. Varian (1999), Information
Rules, Boston, Harvard Business School Press.
Spulber, D. F. (1996), Market Microstructure
and Intermediation, Journal of Economic Pers-
pectives, 10:3, 135-152.
U.S. Department of Commerce (1999), The
Emerging Digital Economy II, Economics and
Statistics Administration, Office of Policy Deve-
lopment, June,
(http://www.ecommerce.gov).
U.S. Department of Commerce (2000), Digital
Economy 2000, June,
(http://www.ecommerce.gov).
22
U.S. Census Bureau of the Department of Com-
merce (2000), Retail E-Commerce Sales in the
First Quarter 2000,
(http://www.census.gov/mrts/www/current.html).
Varian, H. R. and C. Shapiro (1999), conomie
de linformation, De Boeck, Brussels.
White, E. (2000), No Comparison. Shopping
Bots Were Supposed to Unleash Brutal Price
Wars, Why Havent They?, Wall Street Journal,
October 23, 2000, R18.
Williamson, O. E. (1985), The Economic Institu-
tions of Capitalism, The Free Press, New York.